:return: get timing statistics

query the sdf value, semantic label and marching cubes mask for points Args: coord: Nx3 torch tensor, the coordinates of all N (a

assign the queried sdf, semantic label and marching cubes mask back to the 3D grids in the specified bounding box Args: sdf_pred:

use the marching cubes algorithm to get mesh vertices and faces Args: mc_sdf: a*b*c np.array, 3d grids of sign distance values

for each vertex in verts2 find the nearest vertex in verts1 Args: nx3 np.array's scalar truncation_dist: points whose nearest nei

Get the cropped ground truth point cloud according to the intersection of the predicted mesh by different methods Args: file_gt: file

Compute Mesh metrics between prediction and target. Opens the Meshs and runs the metrics Args: file_pred: file path of prediction (sh

get grid query points inside a given bounding box (bbx) Args: bbx: open3d bounding box, in world coordinate system, with unit m

Load a dictionary from a JSON filename. Args: filename: The filename to load from.

read calibration file (with the kitti format) returns -> dict calibration matrices as 4*4 numpy arrays

read pose file (with the kitti format)

Turn a string into a list of string and number chunks. "z23a" -> ["z", 23, "a"]

Computes 2d metrics between two depth maps Args: depth_pred: mxn np.array containing prediction depth_trgt: mxn np.array containi

Load a dictionary from a JSON filename. Args: filename: The filename to load from.